Automated predictive tiered storage system

Patent metadata

Field	Value
Publication number	US-11157194-B2
Application number	US-201815869580-A
Country	US
Kind code	B2
Filing date	Jan 12, 2018
Priority date	Jan 12, 2018
Publication date	Oct 26, 2021
Grant date	Oct 26, 2021

Abstract

The invention relates to a tiered storage system comprising tiers of data storage. The tiered storage system further comprises a processor; and a memory coupled to the processor. The memory comprises instructions which, when executed by the processor, cause the processor to: receive usage data descriptive of usage of memory extents stored by the tiered storage system; identify periodic usage patterns of the memory extents at least partially by calculating a correlation coefficient between the usage data and a predetermined list of conditions; determine a projected data usage for each of the memory extents using the periodic usage patterns, wherein the projected data usage is temporally dependent; sort the memory extents into usage bins according to the projected data usage; and control the tiers of data storage to migrate the at memory extents between the tiers of data storage using temporal changes of the sorting into the usage bins.

First claim

What is claimed is:

1. A tiered storage system comprising:
two or more tiers of data storage;
a processor; and
a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to:
receive usage data, in real time, descriptive of usage of at least one memory extent stored by the tiered storage system;
identify periodic usage patterns of the at least one memory extent at least partially by calculating a correlation coefficient between the usage data and a predetermined list of conditions;
determine a projected data usage for each of the at least one memory extent using the periodic usage patterns, wherein the projected data usage is temporally dependent;
sort the at least one memory extent into usage bins at least partially according to the projected data usage, wherein the sorting into usage bins is temporally dependent, and wherein the commands sent to the two or more tiers of data storage migrate one or more of the at least one memory extent to a different tier; and
control the two or more tiers of data storage to migrate the at least one memory extent between the two or more tiers of data storage using temporal changes of the sorting into the usage bins, wherein a time for the migrating is determined based on a projected data usage.

2. The tiered storage system of claim 1 , wherein the correlation coefficient is a Spearman correlation coefficient, which may be used to determine correlations in both linear and non-linear relations.

3. The tiered storage system of claim 1 , wherein the periodic usage of the at least one memory extent is at least partially identified by calculating a Fourier transform of the usage data.

4. The tiered storage system of claim 3 , wherein the periodic usage of the at least one memory extent is at least partially identified by thresholding the Fourier transform to identify a peak or a region surrounding the peak in the Fourier transformed data.

5. The tiered storage system of claim 1 , wherein the predetermined list of conditions comprises at least one time interval.

6. The tiered storage system of claim 5 , wherein the at least one time intervals is any one of the following: a weekly interval, a bi-weekly interval, a daily interval, a monthly interval, a bimonthly interval, quarterly interval, a yearly interval, and combinations thereof.

7. The tiered storage system of claim 1 , wherein the predetermined list of conditions comprises an external data stream.

8. The tiered storage system of claim 7 , wherein the external data stream comprises any one of the following: predicted weather data, a predicted quantity, a rain probability, an external temperature forecast, a frequency of an internet search term, a social media sentiment index, social media metadata, metadata, stock exchange data, a stock trade volume, a short trade volume, a commodity trade value, a stock index value, and traffic data.

9. The tiered storage system of claim 1 , wherein determination of the projected data usage is performed by linearly extrapolating maximums of the periodic usage patterns, minimums of the periodic usage patterns, and/or portions of the periodic usage patterns.

10. The tiered storage of claim 9 , wherein the linear extrapolation is preformed using a predetermined number of maximums of the periodic usage patterns, minimums of the periodic usage patterns, and/or portions of the periodic usage patterns.

11. The tiered storage of claim 9 , wherein the usage data is a data heat, wherein the projected data usage is a projected data heat, wherein the processor is configured for at least partially sorting the at least one memory extent into usage bins by using a moving heat average for each of the at least one memory extent, wherein a predetermined period before maximums of the projected data usage the projected data usage is substituted for the moving heat average during sorting of the at least one memory extent into usage bins.

12. The tiered storage of claim 11 , wherein the processor is configured for using thresholds of the moving heat average to at least partially determine the sorting of the at least one memory extent into usage bins.

13. The tiered storage of claim 11 , wherein the instructions further cause the processor to receive a current data heat during one of the maximums of the periodic usage pattern, wherein a model correlation is calculated between the current data heat and the projected data heat using a Spearman correlation coefficient, wherein if the model correlation is below a predetermined correlation threshold the performs any one of the following:
disabling substitution of the maximum of the projected data for the moving heat average, wherein the Spearman correlation is used to detect a situation where the projected data usage was not accurate;
repeating the step of identifying periodic usage patterns by calculating the correlation coefficient between the usage data and the predetermined list of conditions;
identifying the step of identifying the period usage patterns using the Fourier transform of the usage data; and
combinations thereof.

14. The tiered storage system of claim 1 , wherein the sorting is at least partially dependent upon a priority assigned to each of the at least one memory extent.

15. The tiered storage system of claim 1 , wherein the usage data is received by any one of the following: logging the usage data, receive usage data via a network connection, and retrieving the usage data from a computer storage.

16. The tiered storage system of claim 1 , wherein the usage data is any one of the following: the data heat, a number of data accesses, and a volume of data access.

17. A method of operating a tiered storage system, wherein the tiered storage system comprises two or more tiers of data storage, wherein the method comprises
receiving usage data, in real time, descriptive of usage of at least one memory extent stored by the tiered storage system;
identifying periodic usage patterns of the at least one memory extent at least partially by calculating a correlation coefficient between the usage data and a predetermined list of conditions;
determining a projected data usage for each of the at least one memory extent using the periodic usage patterns, wherein the projected data usage is temporally dependent;
sorting the at least one memory extent into usage bins at least partially according to the projected data usage, wherein the sorting into usage bins is temporally dependent, and wherein commands sent to the two or more tiers of data storage migrate one or more of the at least one memory extent to a different tier; and
migrating the at least one memory extent between the two or more tiers of data storage using the usage bins, wherein a time for the migrating is determined based on a projected data usage.

18. A computer program product for operating a tiered storage system, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions executable by a processor to cause the tiered storage system to perform a method comprising:
receiving usage data, in real time, descriptive of usage of at least one memory extent stored by the tiered storage system;
identifying periodic usage patterns of the at least one memory extent at least partially by calculating a correlation coefficient between the usage data and a predetermined list of conditions;
determining a projected

Assignees

• IBM

Inventors

• Kimmel Peter R
• Muehge Thorsten
• Rueger Erik

Classifications

• G06F17/156

  using a domain transform, e.g. Fourier transform, polynomial transform, number theoretic transform · CPC title

• G06F3/0649 Primary

  Lifecycle management · CPC title

• G06F3/0604

  Improving or facilitating administration, e.g. storage management · CPC title

• G06F3/0638

  Organizing or formatting or addressing of data · CPC title

• G06F3/0685 Primary

  Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays · CPC title